What causes EOS_SPARK_IMAGE?

A cluster was created on a specific DBR version and was never updated as new LTS versions were released. A long-term job configuration locked to a specific DBR version that has since reached EOS. Automated cluster creation scripts specify a hardcoded DBR version that is now EOS. A Unity Catalog or Photon feature requirement prevented upgrading to a newer DBR version. The upgrade was deferred due to dependency compatibility concerns

How do I fix EOS_SPARK_IMAGE?

Check the Databricks Runtime release notes to find the current Long Term Support (LTS) version. Update the cluster to the latest LTS DBR version. Test your notebooks and jobs on the new DBR version in a staging environment first — look for deprecated API changes. Update any hardcoded DBR version references in cluster creation scripts or Terraform. Schedule regular DBR version reviews to prevent accumulating EOS debt

Medium severitycluster

Power BI Refresh Error:
EOS_SPARK_IMAGE

What does this error mean?

A Databricks cluster is configured to use a Databricks Runtime (DBR) version that has reached End of Support. The cluster will still run, but it no longer receives security patches, bug fixes, or Databricks support assistance.

Common causes

1A cluster was created on a specific DBR version and was never updated as new LTS versions were released
2A long-term job configuration locked to a specific DBR version that has since reached EOS
3Automated cluster creation scripts specify a hardcoded DBR version that is now EOS
4A Unity Catalog or Photon feature requirement prevented upgrading to a newer DBR version
5The upgrade was deferred due to dependency compatibility concerns

How to fix it

1Check the Databricks Runtime release notes to find the current Long Term Support (LTS) version
2Update the cluster to the latest LTS DBR version
3Test your notebooks and jobs on the new DBR version in a staging environment first — look for deprecated API changes
4Update any hardcoded DBR version references in cluster creation scripts or Terraform
5Schedule regular DBR version reviews to prevent accumulating EOS debt

Frequently asked questions

Will an EOS cluster stop working?

Not immediately — EOS means Databricks will no longer support or patch the runtime, but the cluster can still run. However, it will eventually be removed from the workspace and blocked from use.

How often do Databricks Runtime versions reach EOS?

Databricks typically provides 2 years of support for LTS versions. Non-LTS (preview) versions have much shorter support windows. Check the Databricks Runtime support lifecycle page for the exact schedule.